Philosophy
Sources of Uncertainty
Quick fact
In risk analysis, uncertainty is typically divided into two types: aleatory (irreducible randomness) and epistemic (reducible by more knowledge).
Why this is interesting
You make decisions every day, but how often do you consider the hidden sources of uncertainty that could change the outcome?
Read the full explanation
Understanding Sources of Uncertainty
Imagine you're flipping a coin: even with perfect knowledge of physics, you can't predict the outcome because of inherent randomness. That's aleatory uncertainty—it's built into the system. Now imagine trying to guess how many jellybeans are in a jar. You could reduce your uncertainty by counting or weighing the jar—that's epistemic uncertainty, caused by missing information. Sources of uncertainty also include measurement errors (your ruler is imprecise), model simplifications (a map that ignores hills), and future variability (will it rain tomorrow?). By identifying the source, you know whether more data or better models can reduce uncertainty or if you must simply plan for it.
A deeper explanation
The distinction between aleatory and epistemic uncertainty is crucial because it dictates how we handle uncertainty. Aleatory uncertainty stems from inherent variability or randomness in a process—like the exact time of radioactive decay or the next card in a shuffled deck. It cannot be eliminated, only characterized (e.g., by a probability distribution). Epistemic uncertainty, on the other hand, arises from limited knowledge—lack of precise measurements, insufficient data, or incomplete models. It can be reduced by gathering more information or refining models. In practice, most real-world uncertainties are a mix: for example, predicting earthquake damage involves aleatory ground motion variability and epistemic uncertainty about building resistance. Recognizing these sources helps scientists and engineers design experiments, build robust systems, and communicate uncertainty honestly.